Single nucleotide polymorphism profiles of canine T-cell and null-cell lymphomas

Background The histopathological classification of T-cell lymphoma (TCL) in humans has distinctive mutational genotyping that suggests different lymphomagenesis. A similar concept is assumed to be observed in dogs with different TCL phenotypes. Objective This study aimed to identify the previously reported single-nucleotide polymorphisms (SNPs) in both human beings and dogs in canine TCLs and null-cell lymphomas (NCLs) and to design compatible oligonucleotides from each variant based on the multiplex polymerase chain reaction. Methods Genomic DNA was extracted from 68 tumor specimens (62 TCLs and 6 NCLs) and 5 buffy coat samples from dogs with TCL. Four TCL subtypes and NCL were analyzed in 44 SNPs from 21 genes using the MassARRAY. Results The greatest incidences of SNPs observed in all TCL subtypes and NCL ware SATB1 c.1259A > C, KIT c.1275A > G, SEL1L c.2040 + 200C > G, and TP53 c.1024C > T, respectively. Some SNP locations were statistically significant associated with NCL, including MYC p.S75F (p = 0.0003), TP53 p.I149N (p = 0.030), PDCD1 p.F37LX (p = 0.012), and POT1 p.R583* (p = 0.012). Conclusion Each TCL histological subtype and NCL are likely to contain distinctive mutational genetic profiles, which might play a role in lymphoma gene-risk factors and might be useful for selecting therapeutic target drugs for each canine patient.


Introduction
Canine lymphoma can be categorized into B-cell, T-cell, and null-cell (non-B, non-Tcell) lymphomas (1).Regardless of the prevalence of TCL and rare NCLs, a frequent histopathological subtype of TCL is peripheral TCL, not otherwise specified (PTCL-NOS) (2).Other less common subtypes include T-zone lymphoma (TZL), enteropathyassociated T-cell lymphoma (EATCL), and cutaneous T-cell lymphoma (CTCL) (3).CTCL can be either epitheliotropic (ECTCL) or non-epitheliotropic (NECTCL).Canine TCL often has a worse prognosis than B-cell lymphoma (4); however, each TCL subtype often shows a variable prognosis and disease outcome.TZL has a better prognosis because of its indolent disease course (5).Based on the anatomic locations of TCL, patients with hepatosplenic and gastrointestinal lymphoma often have shorter survival times than those with certain forms of multicentric or cutaneous lymphoma (6-10).Moreover, the overall survival time of dogs with ECTCL is shorter than that of dogs with NECTCL (11).Due to their differences in prognosis, a unique therapy for each TCL subtype may be required.Many studies have suggested multidrug chemotherapy for TCL, such as a combination of mechlorethamine, vincristine, prednisone, and procarbazine (12); lomustine, vincristine, procarbazine, and prednisolone (13), or vincristine, L-asparaginase, doxorubicin, cyclophosphamide, actinomycin-D, procarbazine, prednisolone, and lomustine (14), to improve progression-free survival and overall survival time in lymphoma dogs.However, CHOP-based induction (vincristine, cyclophosphamide, doxorubicin, and prednisolone) is still preferred by veterinary oncologists, and lomustine-based protocols are used as rescue treatments (15).Specific molecular signatures of PTCL-NOS in humans can be classified into TBX21 and GATA3 subgroups, which are T helpers (T H )1 and T H 2-cell differentiation regulators, respectively (16).PTCL-NOS-GATA3 frequently exhibited mutations of PTEN and TP53, with co-occurring amplifications of MYC and STAT3.For PTCL-NOS-TBX21, mutations of DNA methylation regulator genes were noted in TET1, TET3, and DNMT3A (17).For human CTCL, specific genetic mutations were involved in T-cell activation (CD28 and RHOA), cell apoptosis (FAS), NF-κB pathways (NFKB2 and STAT5B), chromatin remodeling (DNMT3A and ARD1A), and DNA damage response (TP53 and CDKN2A) (18).Moreover, the loss-of-function mutation of PDCD1-the gene encoding an inhibitory receptor program death protein 1 (PD-1)-was driven by aggressive behavior in CTCL (19).For EATCL, the most frequently mutated genes were associated with the JAK-STAT pathway in SETD2, STAT5B, JAK1, JAK3, and STAT3 (20-22).Hence, these findings are useful for selective potential therapeutic opportunities in each TCL subtype as a single agent or in combination with anti-neoplastic drugs.For example, the NF-κB inhibitor bortezomib and the PI3K-δ,λ inhibitor duvelisib were applied in clinical trials treating PTCL and CTCL, and the overall response rate ranged from 31 to 67% (23,24).
In canine TCLs, Labadie et al. reported SNPs in TZL using genome-wide association mapping (25).Several SNPs of hyaluronidase genes were associated with disease risk for TZL, including SPAM1 K482R , HYALP1 M463T , and HYAL4 G454S, S434F, L378I .Another study investigated somatic mutations in a variable group of TCL with limited subtyping in Golden retrievers and Boxers using whole-exome sequencing (26).The common mutated genes in both breeds were noted on SATB1 c.1259A > G p.Q420R, while PTEN c.975 C > T p.L325 = was found in Boxer-TCLs (26,27).Similar variants using RNA sequencing were found in two dogs with PTCL (27).Three somatic missense mutations were described in canine TCL, including MYC c.185C > T p.Ser62Phe, TP53 c.715G > A p.Arg239Trp, and MET c.3804C > G p.Asp1268Glu (28).Nonetheless, no study has examined novel SNPs associated with human TCL in dogs.Thus, our study aimed to investigate SNP genotyping patterns in canine PTCL, ECTCL, NECTCL, EATCL, and NCL based on the important gene-related lymphoma risks in human and canine TCLs (19, 29-31) and to obligate primer design for multiplex PCR, using the MassARRAY platform (Agena Bioscience, CA, United States).

Study samples and immunohistochemistry
Formalin-fixed paraffin-embedded (FFPE) block archives were retrieved from the Department of Pathology, Faculty of Veterinary Science, Chulalongkorn University, between 2008 and 2021 and from three private veterinary labs (SQ Reference Lab, China; Vet Central Lab, Thailand; and Vet Clinical Center, Thailand) between 2020 and 2021.Only patients who had not previously received chemotherapeutic drugs were enrolled in this study.According to physical examination, abdominal ultrasonography, and buffy-coated smears, all lymphoma dogs were in at least WHO clinical stage III (15).The selected cases included multicentric, cutaneous, and alimentary lymphomas that were recut into 3-μm thickness, stained with hematoxylin and eosin, and immunostained against CD3 (Peter F. Moore, United States), CD20 (ab27093, Abcam, United States), and CD79a (HM57, Abcam) or Pax5 (1EW, Leica, United Kingdom).Immunohistochemistry was performed as described elsewhere (30).Only lymphoma samples that had no immunoreactivity for B-and T-cell markers were further investigated with CD18 (histiocyte marker, Peter F. Moore), CD117 (mast cell marker, MIB1, Dako), and MUM1 (plasma cell marker, ab133590, Abcam) to rule out the possibility of other hematopoietic lineages.Finally, cases that showed negative to specific leucocyte markers but positively immunostained with CD45 (common leucocyte antigen) were assigned as NCL.
To evaluate the effects of missense SATB1 Q420P mutation on programmed death ligand 1 (PD-L1) and missense TP53 I149N mutation on p53, five samples from each group (four SATB1 mutants and one SATB1 wild type vs. two TP53 mutants and three TP53 wild types) were immunolabeled against PD-L1 and p53, respectively.In brief, the tissue section was preheated by citrate buffer pH 6 at 95°C for 20 min and blocked endogenous peroxidase and non-specific binding protein by 0.3% (v/v) H 2 O 2 for 30 min and 5% (w/v) bovine serum albumin for 20 min, respectively.Anti-CD45 antibody (Peter F. Moore) was incubated in a 1:10 dilution at room temperature (RT) for 1.5 h, while PD-L1 (1:100, clone 5D2) (32) and p53 (1: 500, PA5-27822, Invitrogen, United States) were incubated at RT for 2 h.Envision mouse/ rabbit-HRP (Dako, Hilden, Germany) was used as a secondary antibody, and 3,3′-diaminobenzidine was used as a substrate.Histopathological diagnosis was made by a board-certified veterinary pathologist (TK) according to WHO classifications (3).Signalment and clinical history of all selected cases were recorded; however, most cases were lost due to a lack of treatment follow-up.

DNA extraction
A total of 70-100 μm FFPE tissue scrolls were collected into a sterile 1.5 mL microtube.The genomic DNA from each sample was then extracted using a DNeasy Blood and Tissue Kit (Qiagen, Germany).After deparaffinization with xylene and absolute ethanol, the sample was processed according to the manufacturer's instructions.The DNA concentration of each sample was then measured with a NanoDrop Lite Spectrophotometer (Thermo Scientific, United States).The requirement of high-quality DNA for targeted SNP genotyping using the Agena Bioscience MassARRAY system is a 260/280 ratio of >1.8 and a concentration of >10 ng/μL.The dog reference genomes were obtained from Dog10K_ Boxer_Tasha and UMICH_Zoey_3.1. 1MassARRAY was used for genotyping of the 44 targeted SNPs in 21 different genes, as shown in Table 1.The SNP locations were selected based on the previous studies (19,26,27,(29)(30)(31)36). A primer pair for each location was designed by AgenaCx software (Agena Bioscience).First, multiplex PCR primers were amplified in the target regions by using a 5-uL reaction of 0.5 μL of 10X PCR buffer (Agena Bioscience), 2 mM MgCl 2 , 500 μM dNTPs, 0.5 μM of each amplification primer, and 0.2 U DNA polymerase.Second, 1.7 U shrimp alkaline phosphatase (SAP, Agena Bioscience) and 0.17 μL of 10X SAP buffer were added to the first-step PCR product to dephosphorylate residual nucleotides.Third, the iPLEX extension reaction was composed of 0.2 μL of 10X iPLEX buffer (Agena Bioscience), 0.2 μL of 10X dNTP/ ddNTP combination, 0.142 U iPLEX enzyme, and 0.5-1.57μM extend primer mix.These three steps were carried out by T100 thermal cycler (Bio-Rad, United States), and the cycling conditions were followed by Sirivisoot et al. (30).The final extension products were desalted and transferred onto a SpectroChip with an automated nano-dispenser.The different sizes of SNP variations were identified using a matrix-assisted laser desorption/ionization time-of-flight mass spectrometer.Genotyping of each target site was presented as mass spectrum peaks using a MassARRAY analyzer.The percentage of determinate 44 variants (%call rate) for each sample must be achieved by ≥90%.

Statistical analysis
A total of 44 SNP locations were calculated for the correlation with anatomical forms and histological subtypes of TCL/NCL using the chi-square test by GraphPad Prism version 9.0 for macOS (GraphPad Software, United States).
To distinguish a germline from a somatic mutational profile, the authors selected five archival-matched buffy coat samples from PTCL dogs (dog no.7-10 and dog no.14) to assess this SNPs panel compared to TCL specimens.Somatic SNPs were uniquely varied in each dog, as shown in Figure 4.The frequent somatic mutation was seen in TP53 Q342 *(4 out of 4 dogs), while other somatic variants, for instance, MET, TP53, and RHOA were found in 20% (1 out of 5 dogs).
This study investigated the effect of SATB1 Q420P on PD-L1 expression in canine neoplastic T-cells.Four cases of mutant and one case of wild-type SATB1 were immunolocalized for PD-L1.In four mutant dogs, neoplastic T lymphocytes expressed strong cytoplasmic intensity against PD-L1 compared to the wild type (Supplementary Figure S1).Therefore, T-cell anergy and exhaustion might be one of the lymphomageneses in dogs that are related to SATB1 mutation.
Two NCLs having TP53 I149N were evaluated for their p53 expression status.Interestingly, both dogs with TP53 I149N had a loss of p53 expression (<1% of positive nuclei) when compared to three dogs with wild-typed TP53, which showed weak and heterogenous intensity of positive cells <10% (Supplementary Figure S2).

Discussion
In total, 44 targeted SNP panels were designed and evaluated using the MassARRAY in 68 FFPE specimens of canine TCLs and NCLs.The overall pathways of the targeted genes in this study are shown in Figure 5.Most variants were observed in TCL/NCL subtypes except for MYC S75F , LMNB1 S395L , or STAT3 Y640F which were absent in PTCL.The highest frequency of genetic mutations in canine TCL was found at SATB1 Q420P and KIT T425= , respectively, and similar locations were reported in canine nodal B-cell lymphomas (30).
However, the significance of the KIT mutation was synonymous, causing a low impact on carcinogenesis.The variants in hyaluronidase (SPAM1 and ENSCAFG00000024436/HYALP6) and thyroid hormone regulation (DIO2) genes reported in canine TZL (25) were also detected in other histologic TCLs, including PTCL, ECTCL, NECTCL, EATCL, and NCL.Low molecular weight hyaluronan (a byproduct of ligand activation) has pro-inflammatory and pro-oncogenic effects that might be associated with cell proliferation, angiogenesis, and metastasis (37,38).Moreover, three intron variants of SEL1L were present in 34-82% of the TCL/NCL subtypes.Intron mutations impact as a genetic modifier.SEL1L is an unfolded protein response gene that is stimulated during the accumulation of unfolded and misfolded proteins in endoplasmic reticulum stress and plays a role in the protein degradation pathway through the ubiquitinproteosome system (39).In one study, downregulation of SEL1L significantly decreased the expression of TIMP and PTEN involving tumor invasion in human pancreatic cancer (40).Hence, the effect pathway of SEL1L, the mutation function of hyaluronidase genes, and hyaluronan expression levels need further investigation to confirm their contribution to lymphoma pathogenesis in dogs.
Even though POT1 and TRAF3 polymorphisms were not described in four cases of canine TCL (35), we included them in our genotyping panel due to their detection in preliminary samples.A    (46).Another study of mutations affecting the protection of telomere 1 (POT1) explained that POT1 inhibition in CTCL induced telomere fragility, replication fork stalling, and telomere elongation, which led to defective telomere replication during tumorigenesis (47).Thus, a significant mechanism   Bar chart illustrates SNP detection as germline or somatic in five canine TCLs.
Potential pathways of 21 targeted genes contribute to lymphoma development. 10.3389/fvets.2024.1439706 Frontiers in Veterinary Science 09 frontiersin.org of lymphoma development in dogs caused by TRAF3 and POT1 mutations pivotally requires further study.Some SNP locations in RHOA, STAT3, and PDCD1 were frequently described in human TCLs; however, no study has investigated them in dogs with TCL.RHOA encodes a small GTPase that switches signal transduction cascades and promotes cytoskeleton organization, cell migration, and the cell cycle (48).The missense mutations of RHOA p.Asn117Lys and p.Asn117Ile were specifically altered in human CTCLs (36); therefore, we selected these variants and investigated them in four canine TCL/NCL subtypes.Surprisingly, RHOA N117I, N117K was found in all TCL subtypes with the highest frequency in EATCL (4/8, 50%), NCL (3/6, 50%), and ECTCL (10/21, 47%).Another SNP, STAT3 p.Tyr640Phe, was recently discovered to be mutated in human NK-cell lymphoma and CTCL (19, 49).STAT3 is a transcription factor that plays a key role in many cellular processes, including cell growth and apoptosis.Unlike in humans, STAT3 Y640F was rarely observed in canine TCLs/NCLs; it was found in four cases of CTCL (11%), one case of EATCL (12%), and one NCL (16%).Three SNP locations of missense (p.Lys78Arg), stop gained (p.Glu46Ter), and frameshift (p.Phe37LeufsTer35) variants of PDCD1 were found to be loss-of-function mutations in human PTCLs (31).In addition, PD-1 deletions were related to a worse prognosis because they could revoke T-cell exhaustion and drive aggressive behavior in CTCL (19).In dogs, PDCD1 E46 * , K78R, F37LX was detected in NCL (4/6, 66%) and NECTCL (5/15, 33%).Besides, the lack of complete clinical information in the current study concealed the association between the PDCD1 mutation and prognosis.Thus, a prospective study of these SNPs affecting clinical stage, survival time, and disease progression is required to confirm their significance in prognosis.
Our study demonstrated that each canine patient with a specific subtype of TCL showed a different targeted SNP genotype.Hence, specific target therapy based on the SNP information may be advantageous in dogs with a certain type of lymphoma to increase treatment efficiency concurrent with chemotherapy.A study of CTCL in mice illustrated that methyltransferase inhibitors could restore SATB1 function in the Sézary cell line (50), and the usage of this target drug possibly abrogates malignant expansion in CTCL dogs.A STAT3 mutation could cause constitutive activation and might be a possible STAT3 inhibitor target.Napabucasin, a novel STAT3 inhibitor, has shown promising effects in inducing intrinsic and extrinsic apoptosis and downregulating the expression of STAT3 target genes against neoplastic B-cell lineages (51).Furthermore, napabucasin displayed a synergistic effect when administered with doxorubicin in vitro and in vivo experiments and may provide therapeutic implications in human B-cell lymphoma.Immunotherapy of PD-1 is an attractive cancer target.In human clinical trials, PD-1 blockage demonstrated substantial therapeutic activity to treat relapsed/refractory Hodgkin's lymphoma as a single agent and advanced NK/TCL concurrent with chemotherapy (52,53).The effectiveness of novel targeted drugs, particularly somatic SNPs, for treating canine lymphoma is necessitated in preclinical studies.
A small sample size of EATCL and NCL in the present study affected the estimated prevalence of each SNP between the two groups.Moreover, the lack of medical records for each patient hindered the valuable significance of SNPs and prognostic indicators.Another recently critical concern was FFPE-induced mutational artifacts, predominantly excessive T > C mutations, that may be affected in this study.
In conclusion, the MassARRAY platform revealed diverse mutational profiles, exhibiting significant variations across different TCL/NCL subtypes and their anatomical locations.Mutations in SATB1, KIT, SEL1L, and TP53 are frequently observed.Somatic mutations, particularly in TP53, are detected, implying a potential disparity between germline and somatic mutational patterns, although this analysis requires further validation in a larger number of cases and their contribution if they are cancer-associated mutations in dogs.

FIGURE 2
FIGURE 2 Landscape of single-nucleotide polymorphisms in canine T-cell and null-cell lymphomas.Genomic analysis of 21 mutated genes was harbored in 68 dogs with 4 T-cell lymphoma variants: peripheral T-cell lymphoma (PTCL), epitheliotropic cutaneous T-cell lymphoma (ECTCL), non-epitheliotropic cutaneous T-cell lymphoma (NECTCL), enteropathy-associated T-cell lymphoma (EATCL), and null-cell lymphoma (NCL).*No call indicates a failure to detect either a wild-type or mutant peak.

TABLE 1 A
total of 44 specified single nucleotide polymorphisms (SNPs) were selected and investigated in canine T-cell and null-cell lymphomas.